Problem using Copy as cURL (Windows)
I've just updated from Firefox 101.0.1 to Firefox ESR 115.18.0 (I'm still using Windows 7, so can't get a more recent release). I have a program which reads a cURL command from the clipboard then runs cURL repeatedly with slightly modified versions of the command. The command is read from the clipboard rather than pasted into the command prompt window because it may be more than 4096 characters. This no longer works with the updated version of Firefox. Comparing the cURL commands from the two versions of Firefox showed lots of "^" characters had been added. For example "120Months|Yield|RiskScore" had become "120Months^|Yield^|RiskScore". When I edited the cURL command to remove the "^" characters, everything worked like it had before the Firefox update. I suspect this is related to the changes introduced by Bug 1777800. I would like to know whether the problem can be reproduced using up-to-date versions of Firefox and Windows. Starting with web page https://www.trustnet.com/fund/price-performance/t/investment-trusts?norisk=true&PageSize=25 use the network monitor to capture the result of moving to the next page of data, Copy as cURL (Windows) and use the result. Do you get a response containing the data for the next 25 funds or a very small file containing an error message?
所有回覆 (3)
That software isn't a Firefox software? If not then you need to contact their software support to get it update to work with Firefox.
由 markwarner22 於 修改
Perhaps I should have said "I have written a program which reads a cURL command...". I have now updated the program to strip "^" characters which solves my problem. I'm concerned that this change to Copy as cURL will cause problems for other users.
I've discovered that the code which adds the "^" characters (curl.js) has not changed between Firefox release 115 and Firefox release 133 (the latest). So it will be possible to reproduce the problem using an up-to-date version of Firefox. Since the "^" character is an escape character handled by Windows cmd.exe, it is probably not a good idea to include it in the data which cURL sends to a server. I have found several cases where using Copy as cURL (Windows) results in "^" characters being sent to the server. The server then fails to respond or responds with an error. After removing the "^" characters, the server responds with the correct data. Should I report this as a bug using Bugzilla?